Elsevier

Journal of Veterinary Behavior

Volume 7, Issue 5, September–October 2012, Pages 276-282
Journal of Veterinary Behavior

Research
Decreasing dog problem behavior with functional analysis: Linking diagnoses to treatment

https://doi.org/10.1016/j.jveb.2011.10.002Get rights and content

Abstract

Behavioral problems in dogs account for nearly half of the reasons given for relinquishing them to shelters, and thus constitute a significant animal welfare issue. Any successful attempt to manage these problems will require an understanding of the mechanisms that control these behaviors. However, for some of the behavioral problems cited, such as jumping up on people, available treatments are not prescribed after a systematic assessment of the environmental contingencies contributing to the behavior. The current study assesses the use of functional analysis, an established technique for identifying the variables controlling problem behavior in humans, to determine the environmental factors supporting the behavior of jumping up on people in dogs. Statistically significant differences were found in the rate of jumping up behavior across conditions for each dog in the assessment phase. Treatment conditions used the maintaining variable found in the assessment phase. By comparing the rates of jumping up behavior in these conditions, we found the rates to be of lower statistical significance in the treatment condition. Therefore, results show that this methodology is effective in determining the maintaining variables for these individuals, leading to a more precise treatment.

Introduction

About 26% of the dogs living in shelters are relinquished because of behavioral problems (Salman et al., 1998), thereby leading researchers to label these problems a “significant animal welfare issue” (Bennett and Rohlf, 2007, p. 65).

Currently, dog behavioral problem assessment is limited to standardized behavioral testing and questionnaires. Standardized behavioral tests are defined as stimuli serving to elicit behavior during controlled standardized experimental situations where the outcome is compared with that of other individuals placed in the same situations, so as to classify the subject tested (Serpell and Hsu, 2001) (for reviews see Diederich and Giffroy, 2006, Jones and Gosling, 2005). For example, a dog might be considered dog aggressive if it lunges at the dog an experimenter walks by its kennel. Questionnaires are used to attempt to identify the particular problems a dog might possess, (Cottam et al., 2008, Hsu and Serpell, 2003, Segurson et al., 2005), but do not systematically analyze variables that may be maintaining these behaviors.

Many alternative treatments have been proposed for a range of dog problem behaviors. For example, to decrease the dog problem behavior of jumping up on people, within the popular literature, there exist several common techniques, including kneeing the dog in the chest (Koehler, 1996, Pitcairn and Pitcairn, 2005), stepping on a leash to prevent the dog from jumping up (Bridwell, 2007, Lindsay, 2003), and teaching an alternative behavior (such as sitting or laying down) when coming through the door (Coren, 2004, Pitcairn and Pitcairn, 2005, Yin et al., 2008). However, all of these treatments have been proposed in the absence of a standard method with which to empirically diagnose the underlying maintaining variables for the jumping up behavior.

In the field of behavior analysis, Iwata et al., 1994a, Iwata et al., 1994b developed a robust procedure that has proven useful in investigating potential maintaining consequences for problem behavior in a variety of human populations (autistic children, normally functioning children, and low-functioning adults; for a review, see Hanley et al., 2003). This procedure involves the direct observation and repeated measurement of behavior across several conditions that attempt to mimic the possible situations in which the problem behavior is observed by the caregiver. Observation of the environment before the experiment begins helps in evaluating the possible situations within which the problem behavior may occur and how the caregiver reacts in those situations. Validity is assessed by calculating the different rates of responses of the target behavior in each of the conditions. The conditions are repeated until 1 or more are shown to produce the behavior of interest at a high steady rate. Some of the most common conditions are alone, attention, demand, and play. In the alone condition, the individual is left in a barren environment where there are no social or contrived consequences available. The purpose of this condition is to determine whether the individual’s problem behavior is maintained by automatic reinforcement (Vaughan and Michael, 1982). This condition is sometimes replaced with an ignore condition if the individual cannot be left alone or if the behavior requires the presence of another person to be carried out (e.g., hitting a caregiver). In the ignore condition, a caretaker is present but entirely ignores the behavior of the target individual.

The attention condition is conducted to determine whether attention functions as a reinforcing consequence for the individual’s problem behavior. In this condition, the experimenter gives attention when the individual engages in the target behavior. The attention given in this condition should match what the caregiver was doing before the functional analysis was implemented.

The play condition functions as a control procedure in which the problem behavior is not expected to occur. This condition serves as a control because it provides the individual with an environment in which all maintaining variables are freely available (frequent attention, no demands, and tangible items).

The demand condition is conducted to determine whether escape from demands functions as negative reinforcement for the individual’s problem behavior. The demand in this condition should be similar to those delivered in the natural environment (e.g., with a child, the demand might be to ask them to sort or stack items; Fisher et al., 1998).

Although the conditions described previously are the most common because they cover the majority of the environmental variables that have been shown to maintain problem behavior in humans (Carr and LeBlanc, 2003), some researchers have found other variables that may be maintaining the target behavior and have added additional conditions to the assessment. An example is the tangible condition, conducted to determine whether the contingent delivery of a preferred item functions as a reinforcer for the individual’s problem behavior (Hanley et al., 2003). The item is given to the individual only if he or she exhibits the target behavior.

The present study incorporates procedures derived from those described by Iwata et al., 1994a, Iwata et al., 1994b to study variables that maintain the dog behavior of jumping up on humans. This behavior was chosen because it is a common behavioral problem among dogs (Lindsay, 2003, Coren, 2004), that humans complain about, and one that can be potentially dangerous to small children or elderly individuals (Lindsay, 2003) while presenting little threat to the adult experimenters in this study.

Section snippets

Subjects and setting

The subjects recruited for this experiment were 4 dogs aged between 2 and 3 years and belonging to various breeds and both sexes (Table 1). Three dogs completed the experiment. The dogs were volunteered by their primary caregivers. Inclusion criterion: We advertised for dogs who jumped up on people, a main factor was whether or not the dog was engaging in the behavior frequently enough to consider it a problem behavior. Exclusion criterion included dogs that had a history of aggression or were

Results

The Kruskal–Wallis test revealed statistically significant differences in the frequencies for each dog’s jumping up behavior across conditions in the assessment phase (Lola [H = 15.6, df = 4, P < 0.01]; [H = 16.1, df = 4, P < 0.01]; Cole [H = 36.5, df = 4, P < 0.0001]) (Figure 1). Post hoc comparisons confirmed that the tangible condition sustained statistically significant more responding than any other conditions for Pretzel and Lola (P < 0.05, Siegel and Castellan, 1988) (Figure 2). For

Discussion

To date, only 1 previous study has used a functional analysis with a nonhuman animal (Dorey et al., 2009). Dorey et al. (2009) demonstrated that the self-injurious behavior exhibited by an olive baboon (Papio anubis) was maintained by caregiver attention, and successfully extinguished the problem behavior by withholding the attention and giving attention for an appropriate behavior (lip-smacking). The current experiment extends the species used and shows the method’s effectiveness in

Acknowledgments

The authors thank all the dog owners and dogs that participated in this study. The authors are very grateful to Amanda Bosch for her advice throughout the experiment. The authors also thank Amanda Pitt and Alexander Blandina for their help with the data collection.

References (27)

  • N.R. Dorey et al.

    Functional analysis and treatment of self-injury in a captive olive baboon

    J. Appl. Behav. Anal.

    (2009)
  • V.M. Durand et al.

    The Motivation Assessment Scale (MAS) Administration Guide

    (1992)
  • W.W. Fisher et al.

    Functional analysis and treatment of destructive behavior maintained by termination of “don’t” (and symmetrical “DO”) requests

    Appl. Behav. Anal.

    (1998)
  • Cited by (23)

    • A critical review of the evidence for the equivalence of canine and human compulsions

      2021, Applied Animal Behaviour Science
      Citation Excerpt :

      Any comparison with ARBs should investigate whether they serve the same functions. Whilst there are published accounts of functional analyses being applied to other behaviour problems in dogs (Dorey et al., 2011; Winslow et al., 2018; Pfaller-Sadovsky et al., 2019), there are few reports of such investigations of ARBs. Some studies (Moon-Fanelli et al., 2007; Pereira et al., 2010; Moon-Fanelli et al., 2011; Tiira et al., 2012) have used surveys to identify eliciting triggers, but this review could find only one published functional analysis of ARBs (Hall et al., 2015).

    • Perspectives on assessing the emotional behavior of animals with behavior problems

      2017, Current Opinion in Behavioral Sciences
      Citation Excerpt :

      Advocates and proponents include individuals with a background related to the clinical practice of applied behavioral analysis in humans such as Burch and Bailey [20], and Friedman [21], the ‘new wave’ of animal trainers focused on the use of positive reinforcement to shape behavior such as Pryor [22,23] and Ramirez [24,25], as well as individuals within the academic discipline of experimental psychology such as Wynne and co-workers [26•,27,28]. The behaviorist tradition means processes within the central nervous system, like emotional state are not the focus of attention in either diagnosis or management, instead the focus is on the behavior of the individual and the observable events which appear to control it, with treatment coming through the management of reinforcement schedules relating to the behavior [27,28]. Nonetheless, emotional processes are clearly of interest and many advocates of this approach have expressed an enthusiasm for the discipline out of a direct concern for welfare and ethical behavior management [27,29,30].

    • The role of environmental and owner-provided consequences in canine stereotypy and compulsive behavior

      2015, Journal of Veterinary Behavior: Clinical Applications and Research
    • Novel approaches to measuring and changing behavior

      2012, Journal of Veterinary Behavior: Clinical Applications and Research
    • Functional analysis and treatment of problem behavior in 3 animal shelter dogs

      2018, Journal of Veterinary Behavior
      Citation Excerpt :

      Finally, it is important to note that indirect measures have been shown to have low reliability when used to determine behavioral functions (Iwata et al., 2013; Sturmey 1994). Although there are existing studies that demonstrate the effective use of FA in identifying the function of problematic dog behavior, no studies have been conducted using an FA in an animal shelter setting despite its benefits in being applied to such an environment (Dorey et al., 2012). This study is the first to determine the effectiveness of functional analyses with animal shelter dogs and whether implementing function-based treatment improves a dog's perceived adoptability.

    View all citing articles on Scopus
    View full text